2203 - The Role of Lesion-Specific V12Gy as a Novel Predictor of Radionecrosis Following SRS in Multi-Metastasis Cases
Presenter(s)
C. Sclafani1, S. Wadi-Ramahi2, S. R. Ruis2, K. G. Abdullah3, P. Zinn3, S. Sheikh2, C. T. Wilke2, S. Choi2, and A. H. Zureick2; 1University of Pittsburgh School of Medicine, Pittsburgh, PA, 2Department of Radiation Oncology, University of Pittsburgh Medical Center, Pittsburgh, PA, 3Department of Neurological Surgery, University of Pittsburgh Medical Center, Pittsburgh, PA
Purpose/Objective(s): Stereotactic radiosurgery (SRS) for brain metastases has seen an expanding scope of indications, including broadening the upper limit of lesions suitably treated in a single course. However, compared with whole brain radiotherapy, SRS carries an elevated risk of radionecrosis (RN). A commonly cited planning constraint is a global brain or brain-GTV V12Gy under 10cc, under which risk of RN is minimized. Given recent prospective data supporting SRS for up to 20 lesions, it is increasingly difficult to adhere to a constraint validated largely in the limited-metastasis population. Here, we aim to evaluate lesion-specific V12Gy, contrasted with global plan V12Gy, to develop a superior predictor of lesion-specific RN.
Materials/Methods:
We identified consecutive patients receiving linear accelerator-based, single-fraction SRS at our institution (2022-2023) for brain metastases. Inclusion criteria included available RT plans and follow-up MR imaging >3 months. Within the institutional treatment planning system, we generated a 1cm radius ring structure beyond each PTV, as well as a superstructure of all rings. V12Gy was calculated for each lesion’s ring, total ring sum (surrogate for global Brain-PTV), and global Brain-GTV. Chart review included prior CNS RT, systemic agents within 30 days of SRS, vasogenic edema, RN, and management of edema or RN (steroids, bevacizumab, laser interstitial thermal therapy [LITT]). All available MRI scans were personally reviewed for correlation to SRS plans. Receiver Operating Characteristic (ROC) analysis and Generalized Estimating Equations (GEE) were used to compare predictive models and identify optimal dose-volume thresholds while accounting for patient-level clustering.
Results: Thirty-five patients with 138 lesions were identified. Ten lesions in 10 patients developed RN (7.2%). Fourteen patients (40%) had vasogenic edema of irradiated lesions, 13 patients (37%) required corticosteroids, six patients (17%) required bevacizumab, and two patients (6%) underwent LITT. Local V12Gy was a superior predictor of RN compared with global plan Brain-GTV V12Gy (AUC 0.75 vs 0.29). Median local V12Gy for lesions developing RN was 2.21cc compared with 1.02cc (p=0.009). ROC analysis identified 1.80cc as the optimal local V12Gy threshold for predicting RN.
Conclusion: In multi-metastasis SRS, a lesion-specific V12Gy < 1.8cc (within 1cm of PTV) may be a more reliable predictor of RN than global plan metrics. This approach may facilitate safer treatment of a higher number of metastases. A larger cohort will be presented for further validation.